Learners participating in research are considered human subjects and are entitled to same ethical guardrails as patients.
Much of education research may be considered minimal risk and would qualify for IRB exemption, presuming that participation is voluntary and data on whether or not a learner participated is anonymous.
If a learner may not be able to opt out, or if there is a possibility that anonymity is not ensured, this is considered minimal risk and would require IRB review.
For a more in depth analysis of best practices, see the following paper.
One of the pitfalls or stumbling blocks in education research is the availability (or lack of) standardized and validated tools to assess outcome measures. Below is a list of some of these. The true list is extensive and ever-expanding. Helen Chang, our librarian, is an excellent resource in this regard!
Qualitative research methods